added sorting benchmark
Needs RevisionPublic

Authored by fabiank on Aug 27 2019, 7:46 PM.

Details

Reviewers
elvisangelaccio
Group Reviewers
Dolphin
Summary

usr_bin.txt is the content of my /usr/bin as a somewhat realistic input
for the sorting function
citynames.txt is a subsample of city names (as an example of unicode
strings which one might encounter)

Diff Detail

Repository
R318 Dolphin
Branch
benchmark
Lint
No Linters Available
Unit
No Unit Test Coverage
Build Status
Buildable 15709
Build 15727: arc lint + arc unit
fabiank created this revision.Aug 27 2019, 7:46 PM
Restricted Application added a project: Dolphin. · View Herald TranscriptAug 27 2019, 7:46 PM
Restricted Application added a subscriber: kfm-devel. · View Herald Transcript
fabiank requested review of this revision.Aug 27 2019, 7:46 PM

The micro benchmark is rather simplistic, but I hope still better than nothing (and I want it to show that D23507 makes sense)

meven added a subscriber: meven.Jan 13 2020, 6:45 PM

Do you have the results ?

src/tests/CMakeLists.txt
42

Make it optional to build benchmark, perhaps in its own folder.

Sorry, this fell by the wayside. I've had the results attached in a comment in https://phabricator.kde.org/D23507 for my system. I'll make the benchmark optional real soon™…

elvisangelaccio requested changes to this revision.Jan 19 2020, 7:23 PM
elvisangelaccio added inline comments.
src/tests/CMakeLists.txt
42

Why? We already run another benchmark (kfileitemmodelbenchmark) unconditionally.

src/tests/dolphinsortingbenchmark.cpp
30

Use descriptive variable names (e.g. file).

31

Missing const

34–35

Missing braces

56

Not sure what you mean with optimizazion barrier. Can you explain why there is no QVERIFY here?

This revision now requires changes to proceed.Jan 19 2020, 7:23 PM
meven added inline comments.Jan 19 2020, 9:54 PM
src/tests/CMakeLists.txt
42

Morst benchmarks have no assertions.
If a human does not pay attention to the results, running (or buliding them)is pointless, wasting testing and build time.
We should distinguish those kind of test from the rest unless they have assertions of some sort to detect regression in which case their run time should be be reasonable.